The role of prefrontal cortex in sentence comprehension: a rTMS study

Note

The role of the prefrontal cortex in sentence

comprehension: An rTMS study

Rosa Manenti

, Stefano F. Cappa

, Paolo M. Rossini

and Carlo Miniussi

*

a

Cognitive Neuroscience Section, IRCCS S. Giovanni di Dio FBF, Brescia, Italy

b_{Department of Neuroscience, Vita-Salute University and San Raffaele Scientific Institute, Milano, Italy} c_{AFaR Department of Neuroscience, Isola Tiberina and Neurology, University Campus Biomedico, Roma, Italy} d_{Department of Biomedical Sciences and Biotechnologies, University of Brescia, Italy}

a r t i c l e

i n f o

Action editor Yves Von Cramon Published online 23 December 2007 Keywords: Magnetic stimulation Memory resources TMS DLPFC Semantics Syntax

a b s t r a c t

Using repetitive transcranial magnetic stimulation (rTMS), we investigated the role of the left and right dorsolateral prefrontal cortex (DLPFC) in sentence comprehension. Subjects were required to judge which of the two pictures correctly matched the meaning of active and passive semantically reversible sentences (subject–verb–object); the incorrect picture did not match the sentence in term of lexical items (semantic task) or agent–patient struc-ture (syntactic task). The subjects performed the task while a series of magnetic stimuli were applied to the left or right DLPFC. When rTMS was applied to the left DLPFC, the sub-jects’ performance was delayed only for the semantic task, while rTMS applied to the right DLPFC slowed the processing of syntactic information. The results of this experiment pro-vide direct epro-vidence of a double dissociation between the rTMS effects and the type of task, which may reflect a differential hemispheric involvement of working memory resources during sentence comprehension.

ª2007 Elsevier Masson Srl. All rights reserved.

1.

Introduction

The deceptively simple act of understanding the meaning of a common sentence requires a number of cognitive processes. Minimally, these include the analysis of its phonological and syntactic structure, as well as of the meaning of the composing lexical items. Sentence comprehension requires processing a sequence of words, and analyzing their syntactic and the-matic organization in order to create a correct representation

of the entire sentence. This elaboration needs to maintain in an activated state both single word meaning and the syntactic relations between words (Just and Carpenter, 1992). While cur-rent models of language comprehension make diffecur-rent pre-dictions on the proposed time course of syntactic and semantic integration, there is general agreement about the fact that these processes require the temporary storage and manipulation of multiple classes of information (Friederici and Kotz, 2003). Both storage and manipulation are thought

* Corresponding author. Human Physiology, Department of Biomedical Sciences and Biotechnologies, Faculty of Medicine, University of Brescia, Viale Europa 11, 25123 Brescia, Italy.

E-mail address:miniussi@med.unibs.it(C. Miniussi).

a v a i l a b l e a t w w w . s c i e n c e d i r e c t . c o m

j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / c o r t e x

0010-9452/$ – see front matter ª 2007 Elsevier Masson Srl. All rights reserved. doi:10.1016/j.cortex.2006.06.006

to depend upon working memory (WM) resources. Behavioural experiments indicate that the comprehension of syntactically complex sentences is a demanding process, often entailing a high memory load (Turkstra and Holland, 1998; Stromswold et al., 1996; Stowe et al., 1998).

The neural mechanisms involved in sentence comprehen-sion have been investigated by means of lecomprehen-sion studies and of functional neuroimaging. Lesion studies have emphasised the role of damage to perisylvian areas of the left hemisphere in sentence comprehension disorders (Caplan et al., 1996). Nevertheless it has been difficult to ascribe specific functions to discrete regions within this network. The left ventral pre-frontal cortex, in particular Broca’s area, has traditionally been considered to play a central role in syntactic processing (see Grodzinsky, 2000, 2006, for a comprehensive discussion). Imaging studies have provided further evidence for a central role of Broca’s area in syntactic processing (Dapretto and Book-heimer, 1999; Heim et al., 2003; Moro et al., 2001; Caplan et al., 1998, 1999, 2000). During online sentence comprehension, Bro-ca’s area is important not only for syntactic integration, but also for WM mechanisms relevant for language processing (Fiebach et al., 2005). The two factors of syntactic complexity and sentence length have been often confounded in the imag-ing literature. Some investigations have supported the view that the inferior frontal gyrus activation is specific for syntactic processing, while the engagement of the dorsolateral prefron-tal cortex (DLPFC) may reflect the WM load (Caplan et al., 2002; Hashimoto and Sakai, 2002; Walsh and Rushworth, 1999; Mar-tin et al., 2004). A recent fMRI study, which analyzed the areas involved in a sentence judgment task, has underlined the re-cruitment of the dorsal portion of left frontal cortex. Specifi-cally, this area is involved during the processing of syntactic violations associated to a large WM load (Cooke et al., 2006).

It must also be underlined that language comprehension does not involve only the left hemisphere. Less extensive acti-vations have also been observed in right prefrontal cortex dur-ing sentence comprehension (Wartenburger et al., 2004; Moro et al., 2001; Just et al., 1996). This finding may suggest a quanti-tative change of the activation of prefrontal cortex due to a more demanding task (Just et al., 1996) or an engagement of other processing operations, such as visual WM resources, in order to solve a more complex task (Rapp and McCloskey, 1997). Transcranial magnetic stimulation (TMS), a technique that can be used to map the flow of information across different brain regions during the execution of a cognitive task (Walsh and Rushworth, 1999), is another tool which can be applied to the investigation of the neural mechanisms responsible for sentence comprehension. Only one study has investigated the effects of TMS on syntactic processing (Sakai et al., 2002). In that study, a double pulse TMS was delivered to the left in-ferior frontal gyrus or the left middle frontal gyrus, using three possible timings (0, 150, 350 msec after target onset) during a syntactic or semantic decision task. The main finding was that during syntactic decisions there was a shortening of reac-tion times, but only when TMS was applied to the left inferior frontal gyrus 150 msec after target onset. There is limited ad-ditional TMS evidence for a functional role of the prefrontal cortex in language processing. A recent study highlighted a functional dissociation in left inferior frontal cortex (LIFC), with rTMS on posterior LIFC slowing homophone judgments,

while anterior LIFC slowed synonym judgments (Gough et al., 2005). An involvement of left prefrontal cortex in lexi-cal–semantic processing was demonstrated using rTMS (Cappa et al., 2002; Devlin et al., 2003).

More extensive evidence is available that repetitive TMS (rTMS) applied to the prefrontal cortex can interfere with WM tasks. With a few exceptions (Hautzel et al., 2002), most studies support a relationship between, respectively, verbal and visual stimuli, and left and right prefrontal cortex (Manoach et al., 2004; Wendt and Risberg, 2001; Johnson et al., 2003). The activa-tion of BA 46 observed during WM task is lateralized, with a left activation for verbal stimuli and a right activation for visual stimuli (Leung et al., 2002; Rowe and Passingham, 2001; Van der Linden et al., 1999; Stern et al., 2000).

A further functional distinction can be found between the dorsal and the ventral prefrontal cortex within each hemi-sphere (Sala et al., 2003). There is some evidence that the left ventral prefrontal cortex is more relevant for verbal WM, whereas right dorsal prefrontal cortex is more important for the spatial WM (Walter et al., 2003). Recent TMS studies showed that verbal WM is dependent on normal functioning of the middle frontal gyrus bilaterally, suggesting that right and left-sided areas might be involved in parallel processing of different features of stimuli (Mottaghy et al., 2002a, 2003). Moreover the same authors, in a different experiment, showed that the stimulation of the DLPFC impairs perfor-mance in spatial as well as non-spatial tasks. At the same time stimulation of the dorsomedial prefrontal cortex affects only spatial tasks while stimulation of the ventral portion af-fects only non-spatial task (Mottaghy et al., 2002b). A recent review of TMS studies of WM highlighted the involvement of prefrontal cortex both in maintaining the transient patterns of neural activity in other areas that maintain information available online, and in executive processes (Mottaghy, 2006). The aim of this experiment was to evaluate the relative contributions of the left and right DLPFC to sentence compre-hension, by comparing the effects of rTMS applied to these two regions. The experiment involved the presentation of a sentence and two pictures (see Fig. 1). Subjects were re-quired to judge which of the two pictures correctly matched the meaning of a semantically reversible sentence (subject– verb–object) in the active or passive form. In the syntactic con-dition, one picture matched the true meaning of the sentence, while in the ‘‘distracter’’ picture, the subject and the object of the action were reversed. In the semantic condition, the dis-tracter did not correspond to the meaning of the sentence be-cause the subject, object or action did not correspond to the presented lexical items. rTMS was delivered over left or right DLPFC, at a frequency of 10 Hz for a period of 900 msec, start-ing 100 msec after the onset of the display. The paradigm was chosen with the idea to minimise the involvement of the re-tention component of WM, since the information is always present to the subject, and to require the manipulation of the linguistic and pictorial information in order to allow the correct resolution of the matching task. On the basis of lesion and imaging data, we predicted a greater interference with left-sided stimulation, in particular on the syntactic task. This could be attributed to a distant effect in the ventrolateral cortex, responsible for syntactic processing, or to a direct ef-fect on the manipulation component of WM.

2.

Methods

2.1. Subjects

A group of 12 subjects (mean age ¼ 27.33 years; mean edu-cation ¼ 17.67 years) participated in the experiment. All subjects were native Italian speakers and had normal or corrected-to-normal visual acuity. Participants reported being free of neurological disorders or history of seizures. All were right handed, with a mean score on the Edinburgh Handedness Inventory of þ81.33. Participants were informed about the possible risk of rTMS and informed consent was obtained after a safety screening. The experimental methods had ethical approval from the local Human Ethics Committee.

2.2. Stimuli

The stimuli had been selected from the oral and written sen-tence comprehension task in the ‘‘Batteria per l’Analisi dei Deficit Afasici – B.A.D.A.’’ (Miceli et al., 1994). A sentence and two pictures made up each display. All sentences were made up of two noun phrases and one verbal phrase (i.e., ‘‘la mucca segue il cavallo’’ ‘‘the cow follows the horse’’), with a mean length in letters of 26.71 (range 21–37) and a mean length in words of 5.56 (range 5–7). All the experimental sentences and their translations are reported in theAppendix. The stimuli

were presented in black lower-case letters (Times New Roman font size 30) on a white background.

A total of 58 sentence–picture pairs were constructed: 48 for the three experimental blocks (16 for each one), and 10 for the practice session. The subjects’ task was to match the sentence with one of the two pictures. The ‘‘distracter’’ pic-ture could be of two different types: (1) in the semantic condi-tion, ‘‘SEM’’, for the sentence ‘‘the horse is followed by the dog’’ (il cavallo e` seguito dal cane) the wrong picture depicted a cow followed by a dog; (2) in the syntactic condition, in which the sentence was reversible, ‘‘REV’’, for the sentence ‘‘the cow follows the horse’’ (‘‘la mucca segue il cavallo’’) a wrong pic-ture could depict a horse following a cow. In the latter type of trials, morpho-syntactic information is crucial because the subject must correctly comprehend the structure of the sentence to solve the task correctly. Failure to respond cor-rectly to ‘‘REV’’ trials would seemingly indicate a failure to process morpho-syntactic information. In the SEM condition, the crucial information could have been in one of the two noun phrases (e.g., a wrong picture could depict a cow fol-lowed by a dog or a horse folfol-lowed by a cat), or in the verbal phrase (e.g., the wrong picture depicts a horse frightened by a dog). The two pictures used in a trial were always different along one of the two possible dimensions (SEM and REV), and in all the trials there was only one picture that correctly matched the sentence. Each block contained an equal number of SEM and REV trials, presented in a pseudo-random order, and in each condition sentence form (active or passive), Fig. 1 – Presentation of stimuli and rTMS paradigm. rTMS was delivered 100 msec after the onset of the display, which was composed of a sentence and two pictures, and lasted 900 msec (10 stimuli). Participants were instructed to press a left or right key according to the location of the picture consistent with the sentence. The display stayed on the screen until the subject gave a response.

position of the relevant information (the first or the second noun phrase or the verb phrase for the semantic condition), and position of the correct picture (left or right), as well as stimulation site (left, midline sham or right rTMS), were counterbalanced.

2.3. Procedure

Subjects sat in a dimly lit room facing a computer monitor. The stimuli were presented using Presentation software (Ver-sion 0.70,www.neurobs.com) running on a personal computer with a 17-inch screen. At the beginning of each trial, one sen-tence and two pictures were presented. The sensen-tence was lo-cated at the top of the screen above the two pictures, which were on the left and right side of the screen (seeFig. 1). The subjects were asked to judge which of the two pictures was an accurate representation of the sentence. Subjects made a speeded decision via a two-choice button press, using both hands. The visual stimulus stayed on the screen until a re-sponse was given and a blank screen lasting 6000 msec was interspersed between trials. The whole session, including preparation, lasted about 40 min.

2.4. rTMS procedure

rTMS was applied using a Magstim Rapid with a figure-of-eight (double 70 mm) coil which can induce a maximum mag-netic field of 2.2 T at the scalp site. Before the experiment, individual resting motor excitability thresholds of stimulation were determined by stimulating the left motor cortex and in-ducing a contraction evoked by a single TMS pulse in the con-tralateral first interosseus dorsalis muscle. The threshold was defined as the minimum intensity that induced a visible con-traction in the tested muscle, as agreed by two experimenters on at least three out of six trials. The stimulation intensity used during the experiment was set at 90% of each subject’s threshold. The mean stimulation intensity was 56.7% (min. 48%, max. 65%) of the maximum of the stimulator output. During the experiment, rTMS was delivered using a train of 10 pulses with a frequency of 10 Hz (i.e., lasting a total of 900 msec), starting 100 msec after trial onset. On the basis of evoked response studies, this was chosen as an ideal time window for interference with sentence comprehension task (Hald et al., 2006; Eckstein and Friederici, 2005; Haarmann and Cameron, 2005; Haarmann et al., 2005). Moreover, we de-cided to use a 10 Hz frequency to reduce the amount of overall stimuli delivered to each subject. This line of reasoning was based on the aim to keep stimulation on a safe side as much as possible (Wassermann, 1998).

Although individual radiological head images (i.e., mag-netic resonance images – MRIs) were not available for our sub-jects, we localized left and right DLPFCs using the SofTaxic Evolution navigator system (Version 1.0, www.emsmedical. net). This system allows the reconstruction of cerebral cortex in Talairach coordinates, with an accuracy of w1 cm, on the basis of digitized skull landmarks (nasion, inion and two pre-auricular points) from which 50 uniformly distributed points can be mapped out on the scalp (3D Fastrak Polhemus digi-tizer) and then can be related to cerebral anatomy. An estima-tion of the single subject’s cerebral volume is obtained by

‘‘Point-based Warping’’ to a MRI template and a 3D virtual re-construction based on the points recorded from the subject’s scalp.

This method represents a good and cost-effective solution when single subject MRI scans are not available. It should be noted however that, while we can compute with very high precision the location of the coil, this does not strictly imply that we know the precise cerebral areas that are directly influ-enced by the magnetic field. Therefore, we can only assume that we were stimulating the estimated cortical site underly-ing the coil.

Using this system, we localized in each subjects the poste-rior region of BA 46 (Talairach coordinates X ¼ 49, Y ¼ 36, Z ¼ 25, middle frontal gyrus) at about halfway between F3/4 and F7/8, respectively, in the 10/20 EEG system. The subjects wore a close-fitting skullcap on which these positions were reproduced. To stimulate DLPFC, we placed the anterior end of the junction of the two coil wings above this location marked on the skullcap.

The experiment included three blocks corresponding to three stimulation sites: left and right, which were estimated to overlie the left and right DLPFCs, and sham. The stimula-tion site for the sham condistimula-tion was on the midline in the same coronal plane at the same horizontal level as the frontal sites but the coil being perpendicular to the scalp, thus ensur-ing that no effective magnetic stimulation reached the brain during the sham condition. Several approaches could be used to try to ensure that changes in performance are attribut-able specifically to the effects of TMS upon the brain. We de-cided to use both sham stimulation, as a baseline condition, and the stimulation of two homologous areas to compare the effects of rTMS upon different sites. Moreover, in our study we have also taken the approach of observing behaviour across two distinct tasks following the stimulation of one site (e.g.,Beckers and Zeki, 1995). This technique is often used to try to ensure that effects of TMS are specifically due to the modulation of selected brain areas too.

3.

Results

In order to assess the effects of site of stimulation on the two conditions, we analyzed accuracy and reaction time (RT) by means of repeated measures ANOVAs. Since the interaction be-tween condition, site of stimulation and sentence form (active or passive) was not significant [F(2,22)¼2.368; p > .12], the data from the passive and active conditions were combined. Thus, in these analyses, the two factors were condition (semantic, SEM and syntactic, REV) and site of stimulation (left DLPFC, right DLPFC and sham rTMS). Analysis of accuracy [mean ¼ 87.15%; semantic ¼ 87.16% (left ¼ 84.4%, right ¼ 86.5%, sham ¼ 90.6%); syntactic ¼ 86.80% (left ¼ 86.5%, right ¼ 89.6%, sham ¼ 84.4%)] showed no differences between main factors nor significant interactions.

The RT analysis was performed only on correct responses that were less than two standard deviations from the mean RT (the 3% of the responses was eliminated). No significant differences were present for the two main factors [condition: F(1,11) ¼ .233, p > .05; site of stimulation: F(2,22) ¼ .053, p > .05], but there was a significant interaction between them

[F(2,22) ¼ 9.623, p ¼ .001], indicating a differential effects of the site of stimulation in the two conditions. This analysis was followed by planned comparisons (t-tests). As can be seen fromFig. 2, these comparisons confirmed that in the SEM con-dition stimulation of the left DLPFC slowed RTs (2744 msec) more than sham (2509 msec) or right (2429 msec) stimulation [left vs. sham: t(11)¼ 2.111, p ¼ .05; left vs. right: t(11) ¼ 2.956, p ¼ .013]; while in the REV condition stimulation of the right DLPFC slowed RTs (2691 msec) more than left (2519 msec) or sham (2517 msec) stimulation. In this second set of analyses, the comparison between right and sham stimulation was significant [t(11) ¼ 2.449, p ¼ .032] and the difference between left and right stimulation approached significance [t(11) ¼ 1.847, p ¼ .092]. No other effect reached significance.

4.

Discussion

Our study was designed to evaluate the involvement of the left and right DLPFC in understanding semantically reversible ac-tive and passive sentences. The main result is a double disso-ciation of interference effects: rTMS applied to the left DLPFC slowed the performance in the lexical–semantic task; in con-trast, rTMS applied to the right DLPFC slowed the performance in the syntactic task.

These results were not predicted on the basis of the lesion and imaging evidence. A disorder of syntactic comprehension is usually associated to agrammatic Broca’s aphasia, due to le-sions involving the left ventrolateral prefrontal cortex. Lexi-cal–semantic comprehension impairments are typically found in patients with left post-rolandic damage.

This apparent discrepancy may be an example of the fal-lacy of the ‘‘virtual lesion’’ interpretation of TMS effects. The mechanism of action of TMS is to induce electric currents that produce excitation or inhibition in restricted pools of su-perficial cortical neurons underneath the coil. In the last few years, the combined use of TMS and functional imaging tech-niques has allowed the investigation of TMS influence on the underlying cortical areas. In a combined rTMS–PET study,

Mottaghy et al. (2000)demonstrated that rTMS to the right or

left DLPFC can worsen performance in a verbal WM task while inducing significant reductions in rCBF both at the stimulation site and at distant brain regions. These results underlined the ability of rTMS to produce temporary functional modulations of a neuronal network involved in the selected behavioural task that result in a modification of subjects’ performance. Therefore, it is reasonable that the effects highlighted in rTMS studies cannot be totally and directly compared with le-sion studies. Nevertheless, the effects of TMS in the cognitive neuroscience field have often been interpreted as the conse-quence of a transient lesion, which may avoid problems related to cortical plasticity, functional reorganization, indi-vidual differences on subject ability or lesion localization. However, the use of TMS as an interference tool has never pro-duced a clear deficit in the subject performance, like in pa-tients. The type of effect is often related on increased timing in the information processing (e.g., increased reaction time) and, if a reduction of subject performance is recorded, it is most likely explained by the complexity of the processing needed to solve the task; for that reason the effect of TMS may be related to the reduction of the difference between the signal and the noise present in the system. Therefore, TMS can be regarded as an interference method that can in-crease the timing for the information processing or modify criteria for response decision.

With these considerations in mind, we can attempt an in-terpretation of the hemisphere-specific interference effects.

In the semantic condition, the subject had to compare two noun phrases and one verb phrase with their possible pictorial representations. This task requires the retention and the ma-nipulation of lexical–semantic information in WM. Hanten and Martin (2001), Martin and Romani (1994), Martin et al. (1994, 1999)have argued that the span tasks usually adopted to measure verbal WM tap both phonological and semantic retention. The phonological component of span tasks is inde-pendent of the capacity involved in sentence processing, as indicated by the neuropsychological observations of patients with difficulty retaining phonological information who never-theless show preserved sentence comprehension. In contrast, the semantic component plays a crucial role in sentence com-prehension in the maintenance of word meanings prior to their integration with other word meanings. Patients with a semantic retention deficit are impaired in detecting the pres-ence of a semantic anomaly in sentpres-ences with multiple adjec-tives preceding a noun (e.g., ‘‘The wooden big black coat’’) or with several nouns preceding a verb (e.g., ‘‘Priests, melons and cherries can be usually bought in shops’’). The same pa-tients’ performance improved when the adjectives followed the noun or the nouns followed the verb.

A specific role of left prefrontal cortex in the manipulation of semantic information is supported by recent lesion find-ings. A study evaluating a large number of left hemisphere in-jured patients, found that lesions of middle and superior temporal gyrus, mid-frontal cortex in Brodmann’s area 46 and Brodmann’s area 47 of the inferior frontal gyrus affected comprehension performance evaluated with the Curtis– Yamada Comprehensive Language Evaluation (Curtiss and Yamada, 1988), while lesions of Broca’s and Wernicke’s areas did not significantly modify language comprehension evalu-ated with this test (Dronkers and Ogar, 2004).

Fig. 2 – Behavioural results of the experiment. The graph shows reaction times for the three stimulation sites (left, sham and right rTMS), and the two conditions (semantic, SEM and syntactic, REV). As can be seen, rTMS to the left slowed RTs in the SEM condition while rTMS to the right slowed RTs in the REV condition. Vertical bars represent standard errors of the mean.

In the syntactic condition of the experiment, the partici-pants had to create a representation of the meaning of the en-tire sentence, which resulted in the identification of the thematic roles of the lexical items (agent and subject) on the basis of a syntactic analysis, and to compare it with the two presented pictures. This process was prone to interference by right-sided, rather than by left-sided rTMS to the middle fron-tal gyrus. As pointed out in Section1, this is an area which is considered to play an important role in the manipulation of spatial information in WM (Leung et al., 2002; Rowe and Pas-singham, 2001; Van der Linden et al., 1999; Stern et al., 2000; Koch et al., 2005), and thus its role in sentence comprehension was unexpected. The idea of a possible involvement of spatial processes in language understanding can be traced back to Lur-ia’s interpretation of the syndrome of semantic aphasia (Luria and Hutton, 1977), characterized by the defective comprehen-sion of logic-grammatical structures, as due to a disturbances of ‘‘simultaneous, spatial schemes’’. More recently, it has been proposed that the mapping of grammatical categories (subject, object) onto thematic roles (agent, patient) may be based on spatial representations, and thus be particularly sen-sitive to right hemispheric damage. Some clinical observations support this idea (Chatterjee et al., 1995a, 1995b; Warrington, 2000). The hypothesis that the logical structure of a sentence, independent of its syntactic structure, is represented in space

according to a left-to-right ‘‘language line’’ is supported by the recent observation that patients affected by hemi-inatten-tion tend to neglect the posihemi-inatten-tion of the subject noun phrase in sentence presented auditorily. This is in the left part of the tence in the case of actives, but on the right side for passive sen-tences (Rinaldi and Pizzamiglio, 2006).

It is noteworthy that in the case of imaging studies, an in-crease in ‘‘complexity’’ of a sentence, both in terms of syntac-tic structure and of WM requirements, has been associated with more extensive brain activation, extending to the right hemisphere (Just et al., 1996). The recruitment of right-sided areas has been associated with the processing of visuo-spatial content (Drummond et al., 2003; Carpenter et al., 1999).

In conclusion, we propose an interpretation of the ob-served double dissociation in term of interference with specific WM resources, related to each hemisphere. In partic-ular, in the case of the semantic condition, left TMS had an influence on subjects’ performance since it interfered with verbal WM, while in the case of syntactic condition right TMS slowed participants’ performances since it interfered with visuo-spatial memory during sentence–picture match-ing. In particular, since we minimized the involvement of the retention component, these effects may be attributed to an interference with the manipulation of semantic or syntac-tic information.

Appendix

Experimental sentences

Active Passive

Syntactic condition

i gatti inseguono i cani il cavallo e` inseguito dal cane

(the cats chase the dogs) (the horse is chased by the dog)

la macchina segue il camion il nonno e` seguito dal bambino

(the car follows the lorry) (the grandfather is followed by the child)

le mamme baciano i bambini le bambine sono applaudite dal bambino

(the mothers kiss the children) (the girls are applauded by the boy)

le moto hanno superato le macchine la cartella e` coperta dai libri

(the motorcycles outpaced the cars) (the bag is covered by the books)

il cavallo spaventa le bambine le bambine sono precedute dai bambini

(the horse frightens the girls) (the girls are preceded by the boys)

le donne spingono gli uomini i topi sono spaventati dai nonni

(the women push the men) (the mice are frightened by the grandfathers)

le mamme accarezzano le bambine la macchina e` inseguita dalla moto

(the mothers caress the girls) (the car is chased by the motorcycle)

i gatti inseguono i palloni le donne sono inseguite dagli uomini

(the cats chase the balls) (the women are chased by the men)

le moto hanno superato la macchina i topi sono assaliti dai gatti

(the motorcycles exceeded the car) (the mice are assailed by the cats)

il ladro insegue il poliziotto la mamma e` pettinata dal bambino

(the robber chases the policeman) (the mother is combed by the child)

la moto ha superato la macchina i ladri sono inseguiti dai poliziotti

(the motorcycle exceeded the car) (the rubbers are chased by the policemen)

l’autobus segue la macchina le ballerine sono applaudite dal bambino

(the bus follows the car) (the dancers are applauded by the boy)

Semantic condition

la moto insegue la macchina il nonno e` indicato dal bambino

(the motorcycle chases the car) (the grandfather is pointed by the child)

il bambino indica la donna le bambole sono pettinate dalle bambine

(the boy points to the woman) (the dolls are combed by the girls)

r e f e r e n c e s

Beckers G and Zeki S. The consequences of inactivating areas V1 and V5 on visual motion perception. Brain, 118: 49–60, 1995. Caplan D, Alpert N, and Waters G. Effects of syntactic structure

and propositional number on patterns of regional cerebral blood flow. Journal of Cognitive Neuroscience, 10: 541–552, 1998. Caplan D, Alpert N, and Waters G. PET studies of syntactic

processing with auditory sentence presentation. Neuroimage, 9: 343–351, 1999.

Caplan D, Alpert N, Waters G, and Olivieri A. Activation of Broca’s area by syntactic processing under conditions of concurrent articulation. Human Brain Mapping, 9: 65–71, 2000.

Caplan D, Hildebrandt N, and Makris N. Location of lesions in stroke patients with deficits in syntactic processing in sentence comprehension. Brain, 119: 933–949, 1996.

Caplan D, Vijayan S, Kuperberg G, West C, Waters G, Greve D, and Dale AM. Vascular responses to syntactic processing: event-related fMRI study of relative clauses. Human Brain Mapping, 15: 26–38, 2002.

Cappa SF, Sandrini M, Rossini PM, Sosta K, and Miniussi C. The role of the left frontal lobe in action naming: rTMS evidence. Neurology, 59: 720–723, 2002.

Carpenter PA, Just MA, Keller TA, Eddy WF, and Thulborn KR. Time course of fMRI-activation in language and spatial networks during sentence comprehension. Neuroimage, 10: 216–224, 1999.

Chatterjee A, Maher LM, Gonzalez Rothi LJ, and Heilman KM. Asyntactic thematic role assignment: the use of a temporal– spatial strategy. Brain and Language, 49: 125–139, 1995a. Chatterjee A, Maher LM, and Heilman KM. Spatial characteristics

of thematic role representation. Neuropsychologia, 33: 643–648, 1995b.

Cooke A, Grossman M, DeVita C, Gonzalez-Atavales J, Moore P, Chen W, Gee J, and Detre J. Large-scale neural network for sentence processing. Brain and Language, 96: 14–36, 2006. Curtiss S and Yamada J. Curtiss–Yamada Comprehensive

Language Evaluation. Unpublished test, UCLA, 1988. Dapretto M and Bookheimer SY. Form and content: dissociating

syntax and semantics in sentence comprehension. Neuron, 24: 427–432, 1999.

Devlin JT, Matthews PM, and Rushworth MF. Semantic processing in the left inferior prefrontal cortex: a combined functional magnetic resonance imaging and transcranial magnetic stimulation study. Journal of Cognitive Neuroscience, 15: 71–84, 2003.

Dronkers N and Ogar J. Brain areas involved in speech production. Brain, 127: 1461–1462, 2004.

Drummond SP, Brown GG, and JS Salamat. Brain regions involved in simple and complex grammatical transformations. Neuroreport, 14: 1117–1122, 2003.

Eckstein K and Friederici AD. Late interaction of syntactic and prosodic processes in sentence comprehension as revealed by ERPs. Brain Research Cognitive Brain Research, 25: 130–143, 2005.

Fiebach CJ, Schlesewsky M, Lohmann G, von Cramon DY, and Friederici AD. Revisiting the role of Broca’s area in sentence processing: syntactic integration versus syntactic working memory. Human Brain Mapping, 24: 79–91, 2005.

Friederici AD and Kotz SA. The brain basis of syntactic processes: functional imaging and lesion studies. Neuroimage,

20(Suppl. 1): S8–S17, 2003.

Gough PM, Nobre AC, and Devlin JT. Dissociating linguistic processes in the left inferior frontal cortex with transcranial magnetic stimulation. Journal of Neuroscience, 25: 8010–8016, 2005.

Grodzinsky Y. The neurology of syntax: language use without Broca’s area. Behavioral Brain Science, 23: 1–71, 2000. Grodzinsky Y. The language faculty, Broca’s region, and the

mirror system. Cortex, 42: 464–468, 2006.

Haarmann HJ, Ashling GE, Davelaar EJ, and Usher M. Age-related declines in context maintenance and semantic short-term memory. Quarterly Journal of Experimental Psychology – A Human Experimental, 58: 34–53, 2005.

Haarmann HJ and Cameron KA. Active maintenance of sentence meaning in working memory: evidence from EEG coherences. International Journal of Psychophysiology, 57: 115–128, 2005. Hald LA, Bastiaansen MC, and Hagoort P. EEG theta and gamma

responses to semantic violations in online sentence processing. Brain and Language, 96: 90–105, 2006.

Hanten G and Martin RC. A developmental phonological short-term memory deficit: a case study. Brain and Cognition, 45: 164–188, 2001.

Experimental sentences (continued)

Active Passive

(the man combs the girl) (the grandfather is accompanied by the boy)

le bambine pettinano i bambini la macchina e` preceduta dall’autobus

(the girls comb the boys) (the car is preceded by the bus)

i nonni accompagnano la bambina l’autobus e` tirato dal camion

(the grandfathers come with the girl) (the bus is dragged by the lorry)

i cani inseguono i gatti la moto e` stata superata dalla macchina

(the dogs chase the cats) (the motorcycle is exceeded by the car)

il medico saluta il bambino i nonni sono spaventati dai gatti

(the physician says goodbye to the boy) (the grandfathers are frightened by the cats)

il cavallo insegue il cane gli uomini sono spinti dalle donne

(the horse chases the dog) (the men are pushed by the women)

i bambini applaudono le ballerine i conigli sono spaventati dai gatti

(the boys applaud the dancers) (the rabbits are frightened by the cats)

la bambina abbraccia le ballerine il ladro e` inseguito dal poliziotto

(the girl hugs the dancers) (the robber is chased by the policeman)

i pagliacci fischiano al cantante il camion e` spinto dall’autobus

(the clowns clap the singer) (the lorry is pushed by the bus)

le donne inseguono gli uomini il bambino e` seguito dal nonno

Hashimoto R and Sakai KL. Specialization in the left prefrontal cortex for sentence comprehension. Neuron, 35: 589–597, 2002.

Hautzel H, Mottaghy FM, Schmidt D, Zemb M, Shah NJ, Muller-Gartner HW, and Krause BJ. Topographic segregation and convergence of verbal, object, shape and spatial working memory in humans. Neuroscience Letters, 323: 156–160, 2002. Heim S, Opitz B, and Friederici AD. Distributed cortical networks

for syntax processing: Broca’s area as the common denominator. Brain and Language, 85: 402–408, 2003.

Johnson MK, Raye CL, Mitchell KJ, Greene EJ, and Anderson AW. FMRI evidence for an organization of prefrontal cortex by both type of process and type of information. Cerebral Cortex, 13: 265–273, 2003.

Just MA and Carpenter PA. A capacity theory of comprehension: individual differences in working memory. Psychological Review, 99: 122–149, 1992.

Just MA, Carpenter PA, Keller TA, Eddy WF, and Thulborn KR. Brain activation modulated by sentence comprehension. Science, 274: 114–116, 1996.

Koch G, Oliveri M, Torriero S, Carlesimo GA, Turriziani P, and Caltagirone C. rTMS evidence of different delay and decision processes in a fronto-parietal neuronal network activated during spatial working memory. Neuroimage, 24: 34–39, 2005.

Leung HC, Gore JC, and Goldman-Rakic PS. Sustained mnemonic response in the human middle frontal gyrus during on-line storage of spatial memoranda. Journal of Cognitive Neuroscience, 14: 659–671, 2002.

Luria AR and Hutton JT. A modern assessment of the basic forms of aphasia. Brain and Language, 4: 129–151, 1977.

Manoach DS, White NS, Lindgren KA, Heckers S, Coleman MJ, Dubal S, and Holzman PS. Hemispheric specialization of the lateral prefrontal cortex for strategic processing during spatial and shape working memory. Neuroimage, 21: 894–903, 2004. Martin PI, Naeser MA, Theoret H, Tormos JM, Nicholas M,

Kurland J, Fregni F, Seekins H, Doron K, and Pascual-Leone A. Transcranial magnetic stimulation as a complementary treatment for aphasia. Seminars in Speech and Language, 25: 181–191, 2004.

Martin RC, Lesch MF, and Bartha MC. Independence of input and output phonology in word processing and short-term memory. Journal of Memory and Language, 41: 3–29, 1999. Martin RC and Romani C. Verbal working memory and sentence

comprehension: a multiple-components view. Neuropsychology, 8: 506–523, 1994.

Martin RC, Shelton JR, and Yaffee LS. Language processing and working memory: neuropsychological evidence for separate phonological and semantic capacities. Journal of Memory and Language, 33: 83–111, 1994.

Miceli G, Laudanna A, Burani C, and Papasso R. Batteria per l’Analisi dei Deficit Afasici – B.A.D.A. Milano: CEPSAG, Universita` Cattolica del Sacro Cuore, 1994.

Moro A, Tettamanti M, Perani D, Donati C, Cappa SF, and Fazio F. Syntax and the brain: disentangling grammar by selective anomalies. Neuroimage, 13: 110–118, 2001.

Mottaghy FM. Interfering with working memory in humans. Neuroscience, 139: 85–90, 2006.

Mottaghy FM, Doring T, Muller-Gartner HW, Topper R, and Krause BJ. Bilateral parieto-frontal network for verbal working memory: an interference approach using repetitive

transcranial magnetic stimulation (rTMS). European Journal of Neuroscience, 16: 1627–1632, 2002a.

Mottaghy FM, Gangitano M, Krause BJ, and Pascual-Leone A. Chronometry of parietal and prefrontal activations in verbal working memory revealed by transcranial magnetic stimulation. Neuroimage, 18: 565–575, 2003.

Mottaghy FM, Gangitano M, Sparing R, Krause BJ, and Pascual-Leone A. Segregation of areas related to visual working memory in the prefrontal cortex revealed by rTMS. Cerebral Cortex, 12: 369–375, 2002b.

Mottaghy FM, Krause BJ, Kemna LJ, Topper R, Tellmann L, Beu M, Pascual-Leone A, and Muller-Gartner HW. Modulation of the neuronal circuitry subserving working memory in healthy human subjects by repetitive transcranial magnetic stimulation. Neuroscience Letters, 280: 167–170, 2000. Rapp B and McCloskey M. Brain activation and sentence

comprehension. Science, 275: 912–913, 1997.

Rinaldi MC and Pizzamiglio L. When space merges into language. Neuropsychologia, 44: 556–565, 2006.

Rowe JB and Passingham RE. Working memory for location and time: activity in prefrontal area 46 relates to selection rather than maintenance in memory. Neuroimage, 14: 77–86, 2001.

Sakai KL, Noguchi Y, Takeuchi T, and Watanabe E. Selective priming of syntactic processing by event-related transcranial magnetic stimulation of Broca’s area. Neuron, 35: 1177–1182, 2002.

Sala JB, Rama P, and Courtney SM. Functional topography of a distributed neural system for spatial and nonspatial information maintenance in working memory. Neuropsychologia, 41: 341–356, 2003.

Stern CE, Owen AM, Tracey I, Look RB, Rosen BR, and Petrides M. Activity in ventrolateral and mid-dorsolateral prefrontal cortex during nonspatial visual working memory processing: evidence from functional magnetic resonance imaging. Neuroimage, 11: 392–399, 2000.

Stowe LA, Broere CAJ, Paans AMJ, Wijers AA, Mulder G, Vaalburg W, and Zwarts F. Localizing components of a complex task: sentence processing and working memory. Neuroreport, 9: 2995–2999, 1998.

Stromswold K, Caplan D, Alpert N, and Rauch S. Localization of syntactic comprehension by positron emission tomography. Brain and Language, 52: 452–473, 1996.

Turkstra LS and Holland AL. Assessment of syntax after adolescent brain injury: effects of memory on test

performance. Journal of Speech Language and Hearing Research, 41: 137–149, 1998.

Van der Linden M, Collette F, Salmon E, Delfiore G, Degueldre C, Luxen A, and Franck G. The neural correlates of updating information in verbal working memory. Memory, 7: 549–560, 1999.

Walsh V and Rushworth M. A primer of magnetic stimulation as a tool for neuropsychology. Neuropsychologia, 37: 125–135, 1999.

Walter H, Bretschneider V, Gron G, Zurowski B, Wunderlich AP, Tomczak R, and Spitzer M. Evidence for quantitative domain dominance for verbal and spatial working memory in frontal and parietal cortex. Cortex, 39: 897–911, 2003.

Warrington EK. The failure of language comprehension at sentence and phrasal levels in a patient who can speak normally. Cortex, 36: 435–444, 2000.

Wartenburger I, Heekeren HR, Burchert F, Heinemann S, De Bleser R, and Villringer A. Neural correlates of syntactic transformations. Human Brain Mapping, 22: 72–81, 2004. Wassermann EM. Risk and safety of repetitive transcranial

magnetic stimulation: report and suggested guidelines from the International Workshop on the Safety of Repetitive Transcranial Magnetic Stimulation, June 5–7, 1996. Electroencephalography and Clinical Neurophysiology, 108: 1–16, 1998.

Wendt PE and Risberg J. Ethanol reduces rCBF activation of left dorsolateral prefrontal cortex during a verbal fluency task. Brain and Language, 77: 197–215, 2001.